Generally, contextual information relating to business or other entities could be present in structured sources as well as documents. Access control and redaction mechanisms are typically used to protect this information.
On structured data, label-based access control can be used to define the policies that govern access. Similar policies can be enforced on unstructured data, by identifying sensitive data based on the user accessing the document and redacting it. However, keyword searching on such protected data can lead to an inadvertent leakage of information, even if sensitive information is hidden in the results.